Skip to content

AI Check

Chapter: Chapter 7 — Red Teaming with AI & Humans in the Loop Command: /redteam ai-check

Meta-review: mirror, grounding, dissent, sycophancy on AI-assisted analysis.

After model-generated plans or reviews; when analysis feels too agreeable.

Four checks → verdict trust / re-run / reject.

See skill/reference/ai-anti-patterns.md and bias-catalog.md.

Part of The Red Team Handbook, Version 10, maintained by Curate Labs — successor to the defunct UFMCS v9.0 work. See SOURCE.md and NOTICE.md.